EECS 841: Computer Vision

Fall 2008

Location: LEA 1131

Time: MWF 3:00-3:50

Professor: Brian Potetz

Office Hours: Mondays, 4pm - 5pm

Course Description:

Visual information is a major interface between humans and the world around us. Unfortunately, visual data is dauntingly large and complex, and the environmental properties we seek to infer are even more so. As techniques and hardware improve, science has made significant progress in automated scene understanding, and computer vision techniques have been used in robotics, medicine, human/computer interaction, computer graphics, and computer games.

This course offers a broad introduction to the modern methods of computer vision. We will cover the basic problems of the field, its fundamental mathematics, and common tools and solutions. Specific topics will include segmenting images into distinct objects, identifying and locating objects in images, inferring depth from single and multiple images, and other advanced topics. Each student will gain hands-on experience by choosing and developing a project to solve problems from modern computer vision.

Prerequisites:

This course is open to all graduate students. Basic understanding of linear algebra is desirable.

EECS 840 (Image Processing) is not a prerequisite. Background material in image processing will be taught as needed.


Syllabus:

We will cover these topics:

Grading:

Three homework assignments will be given early in the semester.
In each assignment, you will explore the fundamentals of computer vision by programming computer vision algorithms using Matlab.

Each student will work on a final project of their own choosing.
Brief project proposals will be due Monday, Nov 3.
Final projects will be due Monday, Dec 1.
Each student will present their projects to the class in a short, 10 minute presentation,
as well as turn in a written report.

Readings:

There is no required text for this course.
Suggested readings will be supplied with each lecture.

Reccommended reference books:
"Computer Vision, a Modern Approach," Forsyth & Ponce, Prentice Hall, 2003.


Lectures:

Slides in handout formatJust Slides
FridayAug 22Introduction
MondayAug 25Filtering & Edge Detectionpdf
WednesdayAug 27Filtering & Edge Detectionpdf
FridayAug 29Filtering & Edge Detectionpdf
MondaySept 1No class
WednesdaySept 3Matlab Demo
Command History
Matlab Output		pdf
FridaySept 5Edge Detectionpdf
MondaySept 8Edge Detection & Image Pyramidspdf
WednesdaySept 10Hough Transformspdf
FridaySept 12Active Contours (A* Search)
Snakes Matlab Demo		pdf
MondaySept 15Active Contours (Energy Models)
Kass, Witkin, Terzopoulos, "Snakes: Active Contour Models"		pdf
WednesdaySept 17Deformable Templates
Yuille, "Feature Extraction from Faces Using Deformable Templates"		pdf
FridaySept 19Deformable Templates
Felzenszwalb, "Representation and Detection of
Deformable Shapes"		pdf
MondaySept 22Normalized Correlationpdf
WednesdaySept 24Principle Correlation Analysispdf
FridaySept 26PCA & Linear Algebra Reviewpdf
MondaySept 29PCA Applications & Image Formationpdf
WednesdayOct 1Image Formationpdf
FridayOct 3Radiositypdf
MondayOct 6Radiosity & Photometric Stereopdf
WednesdayOct 8Photometric Stereopdf
FridayOct 10Photometric Stereopdf
MondayOct 13Shape From Shading
Zhang, Tsai, Cryer, Shah SFS Survey  		pdf
WednesdayOct 15Project Ideaspdf
FridayOct 17FALL BREAK
MondayOct 20Stereo Reconstruction
Scharstein & Szeliski Stereo Taxonomy		pdf
WednesdayOct 22Midterm Exam
FridayOct 24Stereo Matchingpdf
MondayOct 27Stereo, Motionpdf
WednesdayOct 29Motionpdf
FridayOct 31Motionpdf
MondayNov 3Review of Probabilitypdf
WednesdayNov 5Graphical Models for Computer Visionpdf
FridayNov 7Graphical Models for Computer Visionpdf
MondayNov 10Graphical Models for Computer Visionpdf
WednesdayNov 12Gradient Descent. Graph Cuts.pdf
FridayNov 14Graph Cuts. Gibbs Sampling.pdf
MondayNov 17Gibbs Sampling
pdf
WednesdayNov 19SIFT Features & Object Recognition
Lowe, 1999		pdf
FridayNov 21View Invariant Object Recognition
Savarese, Fei-Fei CVPR 2008		pdf
MondayNov 24Scene Gist. Exemplar-Based Vision.
Oliva, Torralba, 2001
Freeman, Pasztor, Carmichael, 2000		pdf
WednesdayNov 26THANKSGIVING BREAK
FridayNov 28THANKSGIVING BREAK
MondayDec 1Exemplar-Based Vision


Homeworks:

Files associated with homework 1 are located here.
Notice that I've included a Matlab program, PaintEdgeSegments, that I will use to help evaluate your program output. You may find it useful for debugging.

Files associated with homework 2 are located here.

Files associated with homework 3 are located here.